Wi-Fi RSS Indoor Positioning System Using Online Layer Clustering and Weighted DCP-KNN

Abstract

K-nearest neighbors (KNN) methods can be used on indoor positioning system (IPS) based on Wi-Fi fingerprint in the context of internet of things. The positioning of a mobile device (MD) using Wi-Fi technology involves online and offline phases. In this paper, the offline phase includes data collection in WiFi-based Nonintrusive SMS (WinSMS) context, while the online phase involves updating the structure of the collected radio map and online positioning. In online positioning, the proposed Weighted Differential Coordinate Probabilistic-KNN (WDCP-KNN) method based on probabilistic weighting of generalized Reference Points (RPs) and differential coordinates is used. Experiments in a complex indoor environment with real values indicate that the proposed method reduces the positioning error compared to other methods, and is also comparable in terms of computational complexity.